INDEX

Explanations

phrases related to going to specific places

oai_token-act-pair · gpt-3.5-turbo

references to places or locations, particularly "the"

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.11.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

itud

-0.87

thood

-0.78

CLASSIFIED

-0.76

 namely

-0.75

fortunately

-0.69

é¾

-0.69

 followed

-0.69

knit

-0.67

itudes

-0.67

Operation

-0.66

POSITIVE LOGITS

 nearest

1.22

 bathroom

1.18

 restroom

1.13

 cleaners

1.11

 dentist

1.04

 brink

1.00

 afterlife

0.99

 toilet

0.96

 extremes

0.95

gym

0.95

Activations Density 0.141%

phrases related to going to specific places

references to places or locations, particularly "the"

No Comments

No Known Activations